Merge pull request #352 from PESchoenberg/develop

New functions.

grsp0.scm

@@ -94,7 +94,8 @@
 
 (define-module (grsp grsp0)
   #:use-module (grsp grsp3)
-  #:use-module (grsp grsp1)  
+  #:use-module (grsp grsp1)
+  #:use-module (grsp grsp11)  
   #:use-module (ice-9 string-fun)
   #:use-module (ice-9 futures)
   #:use-module (ice-9 binary-ports)
@@ -182,7 +183,7 @@
 	    grsp-sldvls
 	    grsp-clear-on-demand
 	    grsp-ldlc
-	    grsp-padlr
+	    ;;grsp-padlr
 	    grsp-pad-lr
 	    grsp-menuv
 	    grsp-menup
@@ -194,7 +195,10 @@
 	    grsp-movc
 	    plinerc
 	    grsp-file-isolate-name
-	    grsp-pg-psql1))
+	    grsp-pg-psql1
+	    grsp-count-words
+	    grsp-string-lo
+	    grsp-substring-replace))
 
 
 ;;;; pline - Displays string p_s1 p_l1 times in one line at the console.
@@ -2551,7 +2555,7 @@
     res1))
 
 
-;;;; grsp-sldvls - Speparator, Line, display value, line, separator.
+;;;; grsp-sldvls - Separator, Line, display value, line, separator.
 ;; Displays p_n1 blank lines before a spearator defined by p_s2, string
 ;; p_s1 and p_n2 blank lines after p_s1 and separator p_s3.
 ;;
@@ -3014,3 +3018,96 @@
 
     res1))
 
+
+;;;; grsp-count-words - Counts the words in string p_s1 and returns a two
+;; element list:
+;;
+;; - Elem 0: number of words.
+;; - Elem 1: a list containing each word in p_s1 as a separate element.
+;;
+;; Keywords:
+;;
+;; - words, terms
+;;
+;; Parameters:
+;;
+;; - p_s1: string.
+;;
+(define (grsp-count-words p_s1)
+  (let ((res1 '())
+	(l1 '()))
+
+    (set! l1 (string-split p_s1 #\space))
+
+    ;; Compose results.
+    (set! res1 (list (length l1) l1))
+
+    res1))
+
+
+;;;; grsp-string-lo - From string p_s1, leaves only the characters contained in
+;; list p_l1, purging it from everything else.
+;;
+;; Keywords:
+;;
+;; - strings, alphanumeric
+;;
+;; Parameters.
+;;
+;; - p_s1: string.
+;; - p_l1: list of one-character strings ("a" "b", etc-). 
+;;
+(define (grsp-string-lo p_s1 p_l1)
+  (let ((res1 "")
+	(s1 "")
+	(s2 "")
+	(s3 "|"))
+
+    (set! s1 p_s1)
+
+    ;; String loop.
+    (let loop ((j1 0))
+      (if (< j1 (string-length p_s1))
+
+	  (begin (set! s2 (substring s1 j1 (+ j1 1)))
+
+		 ;; If substring is not on list l1, replace it by a
+		 ;; "monentary" substring.
+		 (cond ((equal? (grsp-lal-exists s2 p_l1) #f)
+			(set! s1 (string-replace-substring s1 s2 s3))))
+
+		 (loop (+ j1 1)))))
+
+    ;; Compose results.
+    (set! res1 (string-replace-substring s1 s3 ""))
+
+    res1))
+
+
+;;;; grsp-substring-replace p_s1 - Replaces all p_s2 substrings with substring p_s3
+;; in string p_s1.
+;;
+;; Keywords:
+;;
+;; - replacing, strings
+;;
+;; Parameters:
+;;
+;; - p_s1: string.
+;; - p_s2: string.
+;; - p_s3: string.
+;;
+(define (grsp-substring-replace p_s1 p_s2 p_s3)
+  (let ((res1 "")
+	(b1 #f))
+
+    (set! res1 p_s1)
+
+    (while (equal? b1 #f)
+
+	   (cond ((not (equal? (string-contains res1 p_s2) #f))
+		  (set! res1 (string-replace-substring res1 p_s2 p_s3)))
+		 (else (set! b1 #t))))
+
+    res1))
+

grsp11.scm

@@ -82,7 +82,8 @@
 	    grsp-ly2code
 	    grsp-lal-deletee
 	    grsp-lal-esubstr
-	    grsp-lal-extract-from-lp))
+	    grsp-lal-extract-from-lp
+	    grsp-lal-exists))
 
 
 ;;;; grsp-lal-rel - Replace element in list. Replace element p_j1 of list
@@ -821,7 +822,7 @@
 
 ;;;; grsp-lal-supp - Given list p_l1 with n elements, it returns a
 ;; list that contains one instance per each element value contained in
-;; p_11. That is, it elimitates repeated instances of the elements of
+;; p_11. That is, it eliminates repeated instances of the elements of
 ;; p_l1 and returns its domain subset.
 ;;
 ;; Keywords:
@@ -1262,3 +1263,40 @@
 		 (loop (+ j1 1)))))
 
     res1))
+
+
+;;;; grsp-lal-exists - Finds out if string p_s1 is in list p_l1.
+;;
+;; Keywords:
+;;
+;; - strings, alphanumeric
+;;
+;; Parameters.
+;;
+;; - p_s1: string.
+;; - p_l1: list of strings.
+;;
+;; Output:
+;;
+;; - #t if p_s1 is in the list.
+;; - #f otherwise. 
+;;
+(define (grsp-lal-exists p_s1 p_l1)
+  (let ((res1 #f)
+	(b1 #f)
+	(i1 0))
+
+    ;; List loop.
+    (while (equal? b1 #f)
+
+	   (cond ((equal? i1 (length p_l1))
+		  (set! b1 #t))
+		 ((< i1 (length p_l1))
+
+		  (cond ((equal? p_s1 (list-ref p_l1 i1))
+			 (set! b1 #t)
+			 (set! res1 #t)))))
+
+	   (set! i1 (in i1)))
+
+    res1))

grsp2.scm

@@ -3514,3 +3514,4 @@
 
     res1))
 
+

grsp3.scm

@@ -294,7 +294,8 @@
   #:use-module (grsp grsp4)
   #:use-module (grsp grsp11)  
   #:use-module (ice-9 threads)
-  #:use-module (ice-9 futures) 
+  #:use-module (ice-9 futures)
+  #:use-module (ice-9 string-fun)
   #:export (grsp-lm
 	    grsp-hm
 	    grsp-ln
@@ -517,7 +518,8 @@
 	    grsp-matrix-row-col-setk
 	    grsp-matrix-find-if-prkey-exists
 	    grsp-matrix-row-subexpk
-	    grsp-matrix-row-insert))
+	    grsp-matrix-row-insert
+	    grsp-matrix-tf-idf))
 
 
 ;;;; grsp-lm - Short form of (grsp-matrix-esi 1 p_a1).
@@ -13480,3 +13482,74 @@
 	   (set! res1 (grsp-ms2my res1))))
 
     res1))
+
+
+;;;; grsp-matrix-tf-idf - Creates a matrix for TF-IDF.
+;;
+;; Keywords:
+;;
+;; - frequency, embeddings, words, terms
+;;
+;; Parameters:
+;;
+;; - p_a1: col vector, string, list of documents.
+;;
+;; Notes:
+;;
+;; - https://www.turing.com/kb/guide-on-word-embeddings-in-nlp
+;;
+(define (grsp-matrix-tf-idf p_a1)
+  (let ((res1 0)
+	(s1 "")
+	(s2 "")
+	(td 0)
+	(a1 0)
+	(a2 0)
+	(tn 0)
+	(l1 '("a" "b" "c" "d" "e" "f" "g" "h" "i" "j" "k" "l" "m" "n" "ñ" "o" "p" "q" "r" "s" "t" "u" "v" "w" "x" "y" "z"
+	      "á" "é" "í" "ó" "ú" " "))
+	(l2 '()))
+
+    ;; Create matrices matrices.
+    (set! a1 (grsp-matrix-cpy p_a1))
+    (set! a2 (grsp-matrix-create 0 0 1))
+
+    ;; Add a column to a1 to contain the results of td calculation.
+    (set! a1 (grsp-matrix-subexp a1 0 1))
+
+    ;; Downcase and clean strings of all elements in the first col
+    ;; of a1.
+    (let loop ((i1 (grsp-lm a1)))
+      (if (<= i1 (grsp-hm a1))
+
+	  ;; First we donwcase, trim and clean the string.
+	  (begin (set! s1 (string-downcase (array-ref a1 i1 0)))
+		 (set! s1 (string-trim-both s1))
+		 (set! s1 (grsp-substring-replace s1 "  " ""))		 
+		 (set! s1 (grsp-string-lo s1 l1))
+		 (array-set! a1 s1 i1 0)
+
+		 ;; Calculate the number of terms in each document (td) and
+		 ;; place the results for each document on the second column
+		 ;; of a1.
+		 (array-set a1 (grsp-count-words s1) i1 1)
+
+		 ;; Analize each document and identify each different word.
+		 ;; Place them in a column vector a2 once. If a term already
+		 ;; exists in a2, then do not add a second instance of it.
+		 (set! l2 (string-split s1 #\space))
+		 (set! tn (length l2))
+
+		 (loop (+ i1 1)))))
+
+    ;; Build results matrix.
+    (set! res1 (grsp-matrix-create 0 (grsp-tm a1) (grsp-tm a2)))
+
+    ;; Calculate each termś frequency on each document. Insert the result
+    ;; for each term in each document in res1.
+
+    ;; Eliminate trivial terms.
+
+    ;; Count the number of documents containing each non-trivial term.
+
+    res1))